"the conversation ignores what AI like ChatGPT really is: Not a form of intelligence — which, in and of itself, is almost impossible to define — but rather a large language model that simply scrapes the (inherently flawed) internet and predicts the most likely sequence of words."
This is exactly what I've been saying here since day 1: AI. Does. Not. Exist. It's all just a sham. The proportions of the sham are truly impressive, of course.
@eff @pluralistic @arstechnica Ah, so the whole "AI" thing is transforming from a simple grift to an MLM-like cult. I've been expecting that.
Yes, you can make a computer to tell you that 2*2=5, but it proves nothing.
So when I see how people are surprised that Grok peddles Musk's Nazi propaganda, I can only facepalm.
What did they expect? There were misled by techbros' digital parlor tricks. It's time to wake up and return to the real world where any piece of software does only thing: it turns input into output and nothing more. Apply enough pressure on the input or on the algorithm itself, and you will always get exactly the output you want.
Except, why even bother?
"Machine learning" is a Ponzi scheme, and not just in the traditional, financial sense. Every iteration of it promises something that is technically (scientifically) impossible, and those promises keep piling up.
It's been going on for 65 years. And even though all the previous attempts have failed to deliver (because they were doomed to fail from the start), the grift never dies. Hey, the next generation of ML will be the real actual AI, trust me bro.
No, it won't.
There is no such thing as AI. The so-called "machine learning models" are merely implementations of static algorithms, like any other application.
If you think they're the most complex and sophisticated pieces of software to date, you're so very mistaken. A modern OS kernel is a true technological miracle that contains ideas and snippets of code far more impressive than any glorified database & search engine combo.
Call me crazy, but if you're trying to make a law that bans regulation of a product for 10 years, that product is not something I ever wish to use.
Consider an automation algorithm that fails because its task is mathematically unsolvable due to contradictory definitions or too many variables.
10 years ago, it would've been called an example of overengineering.
Every anthropomorphic feature you ascribe to a machine exists only in your head, only in your imagination. Recognition, generative, hallucinating, and - above all - intelligence. Machines are capable of none of these things.
#AIDoesNotExist #AIHoax
https://www.theregister.com/AMP/2025/04/12/ai_code_suggestions_sabotage_supply_chain/
To this, I can only add that when the ruse fails, the bosses double down by saying the app is guilty and should be held responsible: "It's all the AI's fault!"
>AI has all the hallmarks of a classic pump-and-dump, starting with terminology. AI isn't "artificial" and it's not "intelligent." "Machine learning" doesn't learn. On this week's Trashfuture podcast, they made an excellent (and profane and hilarious) case that ChatGPT is best understood as a sophisticated form of autocomplete – not our new robot overlord.
